Stokes` theorem.

advertisement
[Collect notes on 13.7 and 13.8; collect homework; hand
out practice exam; hand out time sheets for HW #12]
Section 13.8: Stokes’ Theorem
My final exam won’t be as tough as Stokes’ was (see
Stewart’s marginal comment on page 819)! 
Have any of you seen treatments of Stokes’ Theorem etc.
that make more sense than Stewart’s?
..?..
..?..
In future years I might photocopy material from Purcell’s
“Electricity and Magnetism”, which I recall finding more
intuitive than Stewart.
Main ideas? …
..?..
..?..
 Positive orientation of the boundary of a surface: S
 Stokes’ Theorem:
C F  dr = S curl F  dS
provided:
o S is an oriented piecewise-smooth surface
o C (= S) is piecewise-smooth whose orientation
is consistent with that of S
o F is a vector field P,Q,R
o P, Q, and R have continuous partial derivatives
on an open region in R3 that contains S.
 Uses of Stokes’ Theorem:
o Converting a line integral (around a closed curve)
into a surface integral (see Example 1)
o Converting a surface integral into a line integral
(see Example 2)
 Relation between Green’s Theorem and Stokes’
Theorem:
o GT is a special case of ST
o GT is used in the proof of ST
 Analogy between Fundamental Theorem of Calculus,
Fundamental Theorem of Line Integrals, and Stokes
 Curl and circulation:
((curl v)(P))  n = limr0 (C(r) v  dr) / Area(D(r))
where C(r) = D(r), with D(r) the disk with center P,
radius r, and unit normal vector n
 Stokes’ Theorem proves that a vector field F in R3 is
conservative if and only if curl F = 0: in the “hard
direction”, we have
C F  dr = S curl F  dS = S 0  dS = 0
In his discussion of Example 1, Stewart says “Although
there are many surfaces with boundary C, the most
convenient choice is the elliptical region S in the plane y +
z = 2 that is bounded by C.”
But what if we use the surface S consisting of the sides and
bottom of the partial cylinder shown in Figure 3?
What is the contribution to S curl F  dS coming from the
sides of the partial cylinder?
..?..
..?..
Recall that curl F = (1 + 2y) k.
..?..
..?..
Since k is perpendicular to dS, we have k  dS = 0, so there
is no flux of curl F through the sides of S.
What about the bottom of the partial cylinder?
..?..
..?..
We get D (1 + 2y) dA (just as in Stewart’s approach), only
now we didn’t need to us Equation 13.7.10; the unit disk
isn’t being used to parametrize a surface --- it IS the
surface!
Is there a better way to compute D (1 + 2y) dA than
Stewart’s?
..?..
..?..
Look ahead at the answer (π), which gives a clue.
..?..
..?..
Symmetry (specifically, odd symmetry with respect to y)
tells us that D y dA = 0, so D (1 + 2y) dA = D 1 dA =
Area(D) = π.
Can we use symmetry arguments to get a simpler solution
to Stewart’s Example 2? (The fact that the final answer is 0
should make us suspect that there is such a solution!)
..?..
..?..
I wasn’t able to find such a solution, but maybe one of you
can!
Let F = xz i + yz j + xy k as in Example 2, and let S be the
sphere of radius 1 centered at (1,2,3). Find S curl F  dS.
..?..
..?..
Surface integrals over closed surfaces: If S is a closed
surface,
(*) S curl F  dS = 0.
Proof #1: If S is a closed surface, its boundary is the empty
set, so S curl F  dS = S F  dr = 0.
Proof #2: Let C be a piecewise-smooth simple closed curve
that divides S into two pieces S+ and S–, so that C is
positively oriented with respect to S+ and negatively
oriented with respect to S–. [Draw picture.] Then
S curl F  dS = S+ curl F  dS + S– curl F  dS
= C F  dr + –C F  dr = C F  dr – C F  dr = 0.
Hybrid proof: Draw a curve C, but let it shrink down to a
point.
Suppose we have to evaluate C F  dr where C is the unit
circle in the x,y plane, and we suspect that Stokes’ Theorem
would be helpful. What are some surfaces S for which S =
C?
..?..
..?..
The unit disk in the x,y plane; the hemisphere x2+y2+z2=1,
z0; the half-ellipsoid x2+y2+z2/4=1, z0; the paraboloid
z=x2+y2–1, z0; etc.
Download